[00:05] <Kobaz> https://bugs.launchpad.net/bzr/+bug/197426
[00:05] <Kobaz> there's the bug
[00:32] <lifeless> jam: any comments?
[00:32] <lifeless> mwhudson: so, its good you're looking
[00:32] <lifeless> mwhudson: break out the profiler.
[00:33] <lifeless> jam: the Collapsed 1 keys into 1 requests w/ 1 file_ids w/ sizes: [306]
[00:33] <lifeless> jam: is a little noisy
[00:33] <jam> lifeless: take it out
[00:33] <jam> BB:approve
[01:44] <jelmer> lifeless, cool, I'll try it out this week (re subunit)
[01:44] <jelmer> jfroy, pong
[01:47] <jfroy> jelmer: I made a checkout of a very small svn repository (< 10 revisions) that had never been used with bzr-svn today, and when I tried to commit back, I got a "Could not determine revno for <...> because its ancestry shows a ghost at <...>" error.
[01:47] <jfroy> I can send you the repo if you want, it's pretty easy to reproduce.
[01:47] <jfroy> Both bzr and bzr-svn dev.
[01:48] <jelmer> jfroy, If you have an easy way to reproduce it, please file a bug against bzr-svn
[01:48] <jfroy> OK
[01:49] <jfroy> I'll just attach the repository in the bug.
[01:55] <jfroy> https://bugs.launchpad.net/bzr-svn/+bug/339749
[01:57] <jfroy> added repro steps
[02:00] <jelmer> jfroy, works fine here
[02:01] <jfroy> does it?
[02:01] <jfroy> mmmm
[02:01] <jfroy> let me update bzr and bzr-svn
[02:01] <jelmer> jfroy, it works with 0.5.2 as well
[02:02] <jelmer> ah
[02:02] <jelmer> I can now
[02:03] <jelmer> it looks like a regression in bzr
[02:04] <jelmer> 1.12 works fine, bzr.dev breaks
[02:04] <spiv> jelmer: hmm!
[02:05] <jfroy> mm, I still get the error here
[02:05] <jfroy> jelmer: ahh, sorry didn't read your message before that last one
[02:05] <jfroy> OK
[02:06] <jelmer> ah, hmm
[02:06] <jelmer> might be a bug in bzr-svn after all
[02:06] <jelmer> but indeed a regression
[02:19] <fbond> Hm, anyone know why bzr-loom has replaced bzr stat?  I'm not a fan of the changes.
[02:47] <lifeless> fbond: because it tells you about the loom, its been doing it for a while
[02:48] <lifeless> bbiab
[03:21] <mwhudson> lifeless: i think this summarizes my loggerhead performance playing http://pastebin.ubuntu.com/128584/
[04:07] <lifeless> jam: http://pastebin.ubuntu.com/128584/
[04:17] <LaserJock> can you do the equivalent of svn copy with bzr-svn?
[04:20] <lifeless> igc: http://bundlebuggy.aaronbentley.com/project/bzr/request/%3C1236512117.31799.2.camel%40lifeless-64%3E - this recasts the fix you did to get profiling data when ctrl-C is hit to not break layering
[04:20] <lifeless> igc: it will make profiling easier for mwhudson [when things go wrong]
[04:20] <lifeless> igc: please eyeball :)
[04:54] <poolie> lifeless:  that looks pretty sweet
[05:20] <lifeless> poolie: good!
[06:16] <spiv> lifeless: en_AU.UTF-8
[06:20] <lifeless> python -Werror ./bzr selftest test_parents.*test_unicode_symlink -v
[06:21] <poolie> fails with
[06:21] <poolie> bzr: ERROR: exceptions.DeprecationWarning: the sha module is deprecated; use the hashlib module instead
[06:21] <poolie> i need to patch it to ignore that
[06:21] <poolie> there's a jaunty bug open
[06:22] <igc> for me, 4 errors all the same: Unicode unequal comparison failed to convert both arguments to Unicode - interpreting them as being unequal
[06:24] <poolie> DBusException: org.freedesktop.DBus.Error.Spawn.ChildExited: Launch helper exited with unknown return code 0
[06:24] <jam> lifeless: tests fail with 32-bit python 2.5.2 on Hardy
[06:42] <spiv> FWIW I also get that failure, 32-bit python 2.5.2 intrepid.
[06:44] <mwhudson> spiv: which test?
[06:44] <spiv> mwhudson: 17:20 < lifeless> python -Werror ./bzr selftest test_parents.*test_unicode_symlink -v
[06:44] <mwhudson> ah
[08:27] <ja1> mwhudson: lifeless just told me that you need to update your brisbane-core branch
[08:27] <ja1> I did some updates that should make things like "log -v" faster
[08:43]  * mwhudson gives it a go
[08:48] <mwhudson> ja1: https://pastebin.canonical.com/14708/
[08:48] <mwhudson> maybe i need a new groupcompress too?
[08:51] <mwhudson> mm, only a little faster
[10:08] <pysquared> Hello again, could someone help me please.  Been using bzr for a few weeks, have a stable branch and several development ones.
[10:08] <pysquared> I want to add some of the modules created in a dev branch to stable, without creating problems for future merges.  What to do?
[10:41] <bialix> hi, any English people here?
[10:42] <bialix> bzr st now shows the message: pending merge tips: (use -v to see all merge revisions)
[10:42] <bialix> I'm curious why there is "see all merge revisions" but not "see all merged revisions"?
[10:43] <bialix> anybody?
[10:43] <pysquared> Seems a bit quiet in here at the moment.  Sorry, I don't know the answer, not seen it myself yet.
[10:44] <bialix> pysquared: English is foreign language for me. Do you native English speaker?
[10:46] <Peng_> bialix: Maybe it's to help make it clear that the revisions haven't been committed yet.
[10:46] <bialix> what's difference? they're anyway mergeD already
[10:47] <Peng_> Since it says "pending", I guess that's redundant anyway.
[10:47] <bialix> is not "merge" is a verb?
[10:48] <bialix> "all merge revisions" sounds odd for me
[10:48] <Peng_> English is my first language, but I'm kind of terrible at it, so... :P
[10:49] <bialix> heh
[10:49] <bialix> bzr.devs travelling now I guess
[10:49] <Peng_> bialix: I think "merged" would sound more natural, but it doesn't bother me. I think "merge" is sometimes used as a noun.
[10:50] <bialix> "merge revisions"?
[10:50] <bialix> in my language it should be "merged revisions"
[10:50] <bialix> but my language is not English
[10:50] <Peng_> I guess "merge" would be an adjective describing the revisions?
[10:50] <bialix> yep, I guess so
[10:51] <bialix> hence my question
[10:51] <Peng_> "merge revisions" is used like that throughout bzrlib.
[10:53] <bialix> actually 12 times, and half of that count related to status output
[10:54] <bialix> ok, I'm just curious
[10:55] <Peng_> Oh, sorry. I made a mistake when grepping and through it was used about 30 times.
[10:55] <Peng_> Err, s/through/thought/
[10:55] <bialix> does not matter. it's used even for log command
[10:56] <bialix> and there revisions are definitely "merged"
[10:56] <bialix> perhaps it's just my shallow knowledge of English
[11:00]  * SamB wonders why the logs link in the topic doesn't link to this channel's logs
[11:00] <SamB> oh, I guess it's because the logs are broken down first by year ?
[11:01] <SamB> what a strange way to organize logs
[11:02] <Peng_> bialix: I don't have a deep understanding of bzr's internals, so I don't know if the concept of a "merge revision" (a revision that, well, did a merge) makes sense, but I think it does.
[11:02] <Peng_> bialix: Whether it's a better term than "merged revision" in "bzr status" is another matter.
[11:03] <bialix> Peng_: I don't understand
[11:03] <bialix> Peng_: in log.py the "merge revisions" used in the context: "show me merged revisions"
[11:05] <Peng_> bialix: I think the term "merge revisions" makes sense in some situations, but in others it's a mistake.
[11:05] <Peng_> bialix: But I could be way off.
[11:05] <bialix> hi jam
[11:05] <Peng_> It makes sense *to me*, but might not be right.
[11:05] <jam> hi bialix
[11:06] <bialix> I've asked about "merge revisions"
[11:06] <bialix> bonjour vila
[11:06] <bialix> jam: why "merge revisions" but not "merged revisions"?
[11:11] <Peng_> ja1: wb?
[11:12] <ja1> network here is a bit weak
[11:12] <ja1> bialix: "merge revisions" where is that?
[11:12] <Peng_> jam: The output of "bzr status" after "bzr merge", and some other places.
[11:12] <bialix> jam: bzr st now shows the message: pending merge tips: (use -v to see all merge revisions)
[11:12] <bialix> ja1: ^
[11:13] <bialix> ja1: are you in UK at sprint?
[11:13] <poolie> bialix: hello
[11:13] <poolie> we're in brisbane, australia
[11:14] <bialix> hi poolie
[11:14] <bialix> ah
[11:14] <poolie> and john's here, though not here right at the moment
[11:14] <bialix> sorry
[11:14] <poolie> np at all
[11:14] <poolie> we might have a sprint at the time of UDS in late May in Spain
[11:14] <bialix> poolie, I'm curious about "merge revisions" vs. "merged revisions"
[11:14] <poolie> in case you're interested in coming...
[11:14] <poolie> ok
[11:15] <bialix> bzr st now shows the message: pending merge tips: (use -v to see all merge revisions)
[11:15] <bialix> "merge revisions" sounds odd for Russian mind. Is not "merge" is a verb?
[11:15] <ja1> bialix: We're in brisbane, but yes
[11:15] <vila> hi bialix
[11:15] <bialix> vila too?
[11:16] <vila> bialix: yup
[11:16] <bialix> interesting... how do you know that... about Brisbane and Brisbane-core?
[11:16] <bialix> fantastic
[11:17] <vila> but I will go to sleep in a few minutes, jetlag is still hard...
[11:18] <poolie> sleep well
[11:18] <vila> thanks
[11:18] <poolie> bialix: 'merge' is a verb, but in this context i guess it can be used as a noun
[11:18] <poolie> 'commit' is similar
[11:18] <poolie> so is 'diff' for that matter
[11:19] <poolie> i think there's a semantic difference that 'merged revisions' are the ones that were merged in
[11:19] <bialix> poolie: it seems like in the "merge revisions" it's used as adjective
[11:19] <poolie> but the 'merge revision' is the one that did the merging
[11:19] <poolie> this might be a bit subtle or potentially confusing though
[11:19] <poolie> yes, it is an adjective there
[11:20] <bialix> poolie: in log.py there is some places too, where "merge revisions" used in context of "show me merged revisions"
[11:20] <poolie> just in comments or parameters, or also for ui?
[11:21] <bialix> comments/parameters
[11:22] <bialix> int he log help there is "merged revisions" used
[11:22] <bialix> in the log help
[11:22] <poolie> so i haven't checked them, but it'd probably good to change them to 'merged'
[11:23] <bialix> and also "a nested merge revision"
[11:23] <bialix> I'm just trying to understand English better
[11:23] <Peng_> poolie: I agree with you about the semantic difference. Thank you for articulating it well. :)
[11:24] <bialix> then "merge revisions" == "merged but not committed yet", right?
[11:26] <poolie> bialix: it might be 'revisions that did a merge'
[11:26] <poolie> normally we'd say 'pending merge revisions' to mean not committed
[11:27] <bialix> what it means? 'revisions that did a merge'
[11:27] <poolie> revisions that have more than one parent
[11:28] <bialix> ah, ok
[11:32]  * bialix disappears
[11:48] <lifeless> abentley: something like:
[12:57] <thrope> any bzr hosting other than launchpad available yet?
[12:58] <Odd_Bloke> thrope: Any server providing SSH access can do it. :)  But no other centralised hosting that I'm aware of.
[12:59] <lifeless> thrope: savannah, alioth, gna, ...
[12:59] <thrope> anything for private projects?
[12:59] <lifeless> thrope: I don't know; lp can do private projects though
[12:59] <thrope> oh i didnt know
[13:00] <thrope> thanks
[14:02] <SamB> hmm, how come pybaz doesn't have a launchpad project ...
[14:17] <jelmer> SamB, I think it's dead upstream
[14:17] <jelmer> and its source is not in bzr
[14:18] <SamB> jelmer: which upstream?
[14:18] <jelmer> SamB, both baz and pybaz
[14:19] <SamB> pybaz is the Python library for reading baz/tla, isn't it ?
[14:19] <jelmer> yes, but it wraps around the baz command line I think
[14:19] <SamB> OH.
[14:20] <rocky> jelmer: what is the standard way to run the tests on TracBzr, do you know?
[14:21] <jelmer> I just ran "trial tracbzr" IIRC
[14:23] <rocky> trial ?
[14:26] <SamB> EWWW
[14:26] <rocky> oh trial seems to be a twisted test harness
[14:27] <SamB> well, twisted uses trial at any rate
[14:28]  * SamB said EWW because he found out that jelmer was right
[14:29] <jelmer> rocky, trial isn't really tied to twisted though, it's just a test runner
[14:29]  * SamB closes the RFP he had started writing
[14:29] <jelmer> SamB, EWW?
[14:30] <SamB> jelmer: well, I suppose however you implement something like that has got to be disgusting ...
[14:31] <jelmer> SamB, well, it's all legacy code, anyway
[14:31] <rocky> jelmer: i'm trying out nose ... seems to work... but just an fyi i restarted the fix to that _get_weave bug since i know a lot more now and i've done it on a separate branch
[14:31] <rocky> jelmer: i'll set it up for review today
[14:31] <rocky> also gonna add tests
[14:32] <SamB> jelmer: recall that it also does tla
[14:32] <SamB> baz may be legacy, but unfortunately tla isn't :-(
[14:33] <SamB> or should I say "did"?
[14:34] <jelmer> SamB, Are you sure it does tla as well?
[14:34] <SamB> admittedly some people are still using tla only because they haven't figured out how to escape yet
[14:34] <SamB> jelmer: well, okay, supposedly it can import Arch branches
[14:35] <jelmer> SamB, ah, but that's probably because the repo formats from arch and baz are compatible?
[14:35] <SamB> I was kind of hoping pybaz actually understood the format
[14:37] <SamB> anyway, my point is, while baz may be legacy, the format is unfortunately not and tools for importing (or foreign-branching) it would still be useful
[14:38] <SamB> at this point, the easiest way to use an arch repository is probably through git-archimport
[14:38] <SamB> or is it just git-arch ...
[14:40] <SamB> well, that also runs a command-line tool ... maybe it would be good to add tla support to pybaz?
[14:40] <SamB> (that is, git-archimport uses tla internally)
[14:43]  * SamB accidentally finds the GPL'd TLA FAQ in his googles and contemplates a BSD'd TLA FAQ
[14:45] <jelmer> SamB, well, baz and pybaz are still there
[14:45] <jelmer> SamB, and as far as I know they work
[14:45] <jelmer> so there shouldn't be a particular reason to go through git
[14:55] <jelmer> awilkins, hi
[14:55] <jelmer> awilkins, any chance you can file a bug about that branch issue you mentioned yesterday?
[14:55] <jelmer> I can reproduce it
[14:58] <awilkins> jelmer: np.
[15:16] <awilkins> https://bugs.launchpad.net/bzr-svn/+bug/339974
[15:16]  * mtaylor hates to ask this question...
[15:16] <mtaylor> anybody know where I can get recent bzr rpms for SuSE 10.3 ?
[15:18]  * mtaylor cries when people use SuSE
[15:18] <awilkins> God kills an iguana when people use SuSE
[15:19]  * awilkins thinks that might have been funnier if he said "chameleon"
[15:28] <rocky> jelmer: i just proposed my bug branch for merging into trunk for TracBzr
[15:33] <phinze> ugh, there's nothing that hurts more than a mis-resolved conflict in your history
[15:37] <mamruoc> hello people, I can't seem to find how to ignore changes on commited files
[15:40] <mamruoc> ?
[15:52] <rocky> lifeless: hey... if i have a revid and i know what repo it came from, i should be able to figure out which branches in that repo has that rev applied right?
[15:52] <rocky> lifeless: which would NOT be the case if i had the revno (and no branch nick)
[16:01] <Peng_> mamruoc: You can't. You could use "bzr rm --keep whatever.txt" if you want to stop versioning it.
[16:02] <mamruoc> Peng, shouldn't this been a feature bzr should have?
[16:04] <NCommander> Is it possible to take a CVS checkout, and then import the versions that are present in that checkout into CVS?
[16:05] <Peng_> mamruoc: Yes, but it doesn't.
[16:06] <Peng_> NCommander: There are, um, CVS to bzr converters that can work remotely..
[16:06] <NCommander> Peng, I'm not being clear
[16:06] <NCommander> I have a working tree, from abotu six months ago
[16:06] <NCommander> I want to convert history from that point into bazaar, but nothing from the future
[16:07] <phinze> hmm i'm having trouble pushing to a new lp branch... bzr: ERROR: File exists: '/srv/bazaar.launchpad.net/push-branches/00/00/97/8a/.bzr'
[16:07] <phinze> any clues on how to fix this?
[16:08] <mamruoc> Peng, are the any ongoing work for this?
[16:10] <phinze> hmm i had registered branch prior to pushing, and deleting the branch in lp and just pushing seems to have worked
[16:14] <phinze> for the convenience of anyone looking for the code until the server goes back up, and i'm sure for the annoyance of sid3w1nder...
[16:14] <phinze> https://launchpad.net/bitlbee-snapshot
[16:15] <phinze> :)
[16:17] <Peng_> mamruoc: I don't know of any.
[16:17] <phinze> oh crap
[16:17] <phinze> wrong channel
[16:17] <phinze> why do i keep the two "b" channels next to each other
[16:17] <Peng_> NCommander: Oh. I dunno. Worst case, convert the whole thing, then "bzr uncommit" or "bzr branch" to get rid of it.
[16:18] <NCommander> Peng, oh yay, uncommitting a year of commits -_-;
[16:18] <NCommander> Is there a good guide on how I go about doing this? (or should I just let Launchpad's CVS importer work)
[16:18] <Peng_> NCommander: uncommit can handle multiple revisions at a time.
[16:19] <Peng_> Sorry, I dunno.
[16:58] <rocky> hey ... i'm working on code in a branch ... i did a merge from another branch and forgot to commit it ... i since then changed more code and was about to commit when i realized my changes are messed up with the merge  commit ... is there anyway to separate the merge changes and my local changes at this poi8nt?
[17:11] <Peng_> rocky: You could use "bzr shelve" to temporarily get your new changes out of the way.
[17:15] <Tak> or unmerge, commit, remerge(, commit)?
[17:17] <rocky> Tak: ah that's what i was looking for
[17:17] <rocky> thx
[17:19] <Peng_> Oh, you had committed the merge? Sorry I misunderstood.
[17:20] <rocky> no i hadn't committed the merge
[17:20] <rocky> which means i misunderstood ;)
[17:21] <Peng_> Oh.
[17:21] <rocky> does anyone know if branch.revision_history() ever changed such that once it listed individual merge items VS now just returning the revision belonging to the complete merge?
[17:44] <rockstar> Hm, the last Planet Bazaar post was on March 3.  Surely something is wrong.
[17:47] <Odd_Bloke> rockstar: "Last updated: March 03, 2009 11:06 AM" <-- looks like it
[17:59] <jfroy> with all the talks about changing the default branch format, the thought occurred to me: would it be useful to have a configuration variable for the default repository / branch format?
[18:38] <Peng_> jfroy: I'm not sure it's worth the confusion it could cause. Since DVCSes are all about collaborating with other people, I think it's best for everybody to be in the same default format boat.
[18:39] <Peng_> jfroy: OTOH, always typing "bzr init-repo --1.9" is getting old... :P
[18:39] <jfroy> Peng_: well they are changing it to 1.9 :p
[18:39] <Peng_> ...Good point.
[18:40] <jfroy> I had in mind my workflow, which involves creating repos with 1.9-rich-root all the time (I interface with subversion repos. a lot)
[18:40] <Peng_> jfroy: Yeah, but if you set your default to 1.9-rich-root, it's easier to accidentally create rich-root repos for projects that shouldn't have them.
[18:41] <jfroy> arguably bazaar shouldn't have the rich root thing. It's a screw up from the early days :/
[18:43] <Peng_> Yep. The problem now is dealing with that mistake.
[18:49] <jfroy> It should have been dealt with a long time ago...
[18:49] <jfroy> The longer it stays around, the harder it will be to fix it, IMHO.
[18:49] <jfroy> Anyways :/
[18:51] <Peng_> jfroy: The current (tentative?) plan is to do it with the brisbane-core/CHK formats. Upgrading them requires reserializing inventories(?) anyway, so it's a good time to do it.
[18:51] <Peng_> s/them/to them/
[18:51] <jfroy> Interesting. brisbane-core will be *yum*. The speed, oh my!
[18:52] <jfroy> Well, hopefully.
[19:13] <rocky> so... what's the problem with using --1.9-rich-root on a repo that shouldn't have them?
[19:23] <Peng_> rocky: If anyone else has a non-rich-root copy, they have to upgrade before they can pull from you.
[19:23] <Peng_> rocky: Unless the project is willing to upgrade to rich-roots, your branches will be effectively unmergeable.
[19:24] <jfroy> https://bugs.launchpad.net/bzr/+bug/340108 :/
[19:25] <thrope> I have a bzr-svn question (ping: jelmer) - I have two diverged branches, craeted with bzr with nothing to do with svn... they have a shared history up to revision 24 or so. I would like to push these both into a svn repository... at the moment I just tried svn-push (push caused a traceback) each one to two different dirs, but all the common commits appear twice
[19:26] <thrope> is there a way to push them while keeping the branch relationship in svn terms (ie have one as a copy of the other)
[19:26] <jelmer> thrope, if you got a traceback from svn-push, please file a bug
[19:26] <thrope> jelmer: from normal push, not svn push
[19:27] <jelmer> thrope, still, please file a bug
[19:27] <thrope> also: I always get upgrade to subversion 1.5 message although I am using 1.5.5 on both client and server
[19:27] <jelmer> thrope, that's fixed in bzr-svn trunk
[19:29] <thrope> the error is:bzr: ERROR: exceptions.NotImplementedError: <bound method SvnBranch.push of SvnBranch('file:///Users/robince/tmp/test/amaripool/analysis_version')>
[19:29] <rindolf> Hi all.
[19:30] <jfroy> jelmer: https://bugs.launchpad.net/bzr-svn/+bug/340110
[19:30] <rindolf> Where can I find the bzr-svn-0.5 branch? http://bazaar-vcs.org/BzrForeignBranches/Subversion#development seems to be empty.
[19:30] <jfroy> rindolf: http://people.samba.org/bzr/jelmer/bzr-svn/0.5
[19:30] <Peng_> rindolf: bzr branch http://people.samba.org/bzr/jelmer/bzr-svn/0.5/
[19:31] <Peng_> rindolf: It's not actually empty.
[19:31] <jfroy> it's a branch :p
[19:31] <rindolf> Peng_: thanks.
[19:32] <rindolf> Well, it would be useful to make the web-service put some nice information there.
[19:32] <rindolf> Peng_: with the command you gave me, etc.
[19:33] <Peng_> rindolf: It's hardly a "web-service". The HTTP server is just serving a bunch of opaque files.
[19:33] <rindolf> Peng_: then why is it empty?
[19:34] <Peng_> rindolf: Because the files are in the directory ".bzr", and Apache's directory listings don't show dotfiles by default.
[19:34] <rindolf> 11464kB - holy shit.
[19:34] <rindolf> Do you store .iso's there?
[19:37] <hmeland> rindolf: No, but you now have a local copy of all the 3861 (or so) revisions that make up bzr-svn's history.
[19:37] <rindolf> hmeland: I see.
[19:38] <hmeland> If you just want the current code, you can use "bzr checkout --lightweight http://...".
[19:38] <rindolf> $ ls -l bzr-svn-0.5.2.tar.gz
[19:38] <rindolf> -rw-r--r-- 1 shlomi shlomi 232794 2009-02-18 10:09 bzr-svn-0.5.2.tar.gz
[19:39] <Peng_> Though you've already downloaded it, so it would be a bit silly to delete it, unless you really need the disk space...
[19:46] <vadi2> Hi, I have a question. I need to make a bzr branch that is able to update stuff from a git repository, but ignores the stuff that comes from git for revisioning
[19:46] <vadi2> I found this: http://bazaar-vcs.org/BzrForeignBranches/Git, which depends on Dulwich, and has a Ubuntu PPA, but the ppa only has 32bit packages
[19:47] <vadi2> Is Dulwich available for 64bit?
[19:47] <mwhudson> you can install it from source, it seems to work
[19:47] <vadi2> Ok
[19:48] <rocky> rindolf: fwiw, i'm working on ClueBzrServer which will allow for http-based browsing (akin to mod_svn) for current revisions of files from bzr
[19:48] <rindolf> rocky: ok, thanks.
[19:48] <mwhudson> (though tests don't pass on 64 bit)
[19:55] <vadi2> how would I install dulwich? I'm not sure how to make setup.py work
[19:55] <vadi2> ahh nm
[19:56] <rindolf> OK, https://bugs.launchpad.net/bzr-svn/+bug/311712 - I was able to build an .rpm out of subvertpy and bzr-svn.
[19:56] <rindolf> Thanks everybody.
[19:58] <vadi2> Another question... I have a -lot- of bzr branches from various projects. Is there any tool to check which are outdated?
[20:00] <mwhudson> vadi2: there is a plugin that can tell you which are merged into some other branch
[20:01] <vadi2> mm
[20:02] <mwhudson> vadi2: https://edge.launchpad.net/bzr-removable
[20:02] <vadi2> I see
[20:12] <jelmer> mwhudson, tests do pass on 64bit these days
[20:13] <mwhudson> jelmer: oh right
[20:14] <mwhudson> jelmer: want to mark https://bugs.edge.launchpad.net/dulwich/+bug/337483 as fixed then?
[20:14] <jelmer> mwhudson, we need a IRC interface to Launchpad :-)
[20:14] <mwhudson> indeed
[20:14] <jelmer> ubottu, bug 337483: status fixreleased
[20:25] <vadi2> I installed bzr-git as the help file says (move to ~/.bazaar/plugins and rename to git), but now it says this: "Unable to load plugin 'git' from '/home/vadi/.bazaar/plugins'"
[20:25] <luca> excuse me I'm new and I have a question
[20:26] <Peng_> vadi2: There should be details in ~/.bzr.log
[20:26] <luca> I created a branch on launchpad, now I made some changes to the code that I can see on bzr diff
[20:26] <luca> how do I commit the changes to that branch?
[20:26] <Peng_> luca: "bzr commit". You should read the tutorial.
[20:26] <Peng_> Or something.
[20:27] <luca> but if I write commit, is it sure that the change goes to the launchpad branch?
[20:27] <vadi2> "from bzrlib.foreign import foreign_vcs_registry
[20:27] <vadi2> ImportError: No module named foreign
[20:27] <vadi2> "
[20:29] <Peng_> vadi2: Your bzr is too old.
[20:29] <Peng_> vadi2: Or it's broken. I dunno.
[20:29] <vadi2> i'm on ubuntu 8.10
[20:29] <Peng_> Ok.
[20:30] <Peng_> vadi2: And your bzr is too old.
[20:30] <Peng_> vadi2: You could use https://launchpad.net/~bzr/archive/+ppa
[20:30] <Peng_> (if that's the right link)
[20:30] <jelmer> vadi2: if you're trying out bzr-git, you'll need bzr.dev
[20:31] <Peng_> Oh.
[20:32] <vadi2> where can I obtain that?
[20:32] <Peng_> bzr branch http://bazaar-vcs.org/bzr/bzr.dev/
[20:32] <Peng_> (if that's the right link)
[20:32] <jelmer> Peng_, yes, it is
[20:33] <vadi2> ' ERROR: Unknown repository format: 'Bazaar RepositoryFormatKnitPack6 (bzr 1.9)\n'
[20:33] <vadi2> '
[20:33] <Peng_> Heh, and the recent format upgrade finally hurts someone.
[20:34] <jelmer> Peng_, I think it's already hurt more people
[20:34] <jelmer> Peng_, lenny has 1.5
[20:34] <Peng_> Nice.
[20:47] <vadi2> okay, bzr.dev installed
[20:57] <thecookie> I think I just set up something similar to http://doc.bazaar-vcs.org/bzr.dev/en/user-guide/index.html#team-collaboration-distributed-style - The server hosting the main branch will also host a web page of the repo. Is there something wrong with using the repo as the document root of a web server, is there a better way to do it?
[20:57] <vadi2> Sorry... I tried git-import, and get this: "bzr: ERROR: No module named commit
[20:58] <vadi2> You may need to install this Python library separately.
[20:58] <vadi2> "
[20:59] <Peng_> vadi2: I think that's a dulwich bug. jelmer, you deleted dulwich/commit.py but repo.py still imports it.
[20:59] <jelmer> vadi2, do you have the latest dulwich (from trunk) and bzr-git (from my branch)  ?
[21:00] <vadi2> I have both latest from their main branches
[21:00] <vadi2> so maybe not bzr-git from your branch
[21:01] <Peng_> ...And then it imports object.Commit, overwriting commit.Commit...
[21:01] <Peng_> Err, objects.Commit. Plural.
[21:02] <jelmer> Peng_, vadi2: please try again
[21:02] <vadi2> jelmer: bzr-git from trunk is newer than your branch
[21:02] <vadi2> try what?
[21:02] <jelmer> vadi2, with the latest revision from my branch
[21:02] <jelmer> vadi2, my branch is at http://people.samba.org/bzr/jelmer/dulwich/trunk
[21:02] <Peng_> jelmer: Oh, perfect.
[21:03] <Peng_> :)
[21:04] <vadi2> getting your branch made it crash
[21:04] <vadi2> http://paste.pocoo.org/show/107143/
[21:05] <vadi2> oh, hm. But I do have dulwich installed.
[21:06] <vadi2> what is the proper procedure for installing it? I used ./setup.py build, sudo ./setup.py install
[21:07] <jelmer> vadi2, for dulwich, that should work
[21:08] <vadi2> yeah, but I tried getting your branch, and got this: http://paste.pocoo.org/show/107143/
[21:08] <vadi2> it says I need to install dulwich and then crashed
[21:18] <jelmer> lifeless, what should I replace multiply_tests_from_modules with?
[21:20] <vadi2> Is there a command to delete all unversioned files?
[21:22] <jelmer> vadi2, bzr clean-tree
[21:23] <vadi2> thank you
[22:04] <thecookie> Anyone know if the bzrtools package includes the push-and-update plugin?
[22:05] <Peng_> thecookie: It doesn't.
[22:05] <Peng_> Last I saw anyway...
[22:10] <meoblast001> hi
[22:16] <meoblast001> how do i revert back to a specific revision?
[22:17] <Peng_> meoblast001: The working tree, or do you want to remove revisions from history as well?
[22:17] <Peng_> meoblast001: For the former, "bzr revert -r 123".
[22:18] <CBro2007> how do I install bzr on centos?
[22:18] <meoblast001> Peng_: 123? you mean the revision number?
[22:18] <CBro2007> is there a yum install for it?
[22:19] <Peng_> meoblast001: Yes.
[22:19] <meoblast001> k thanks
[22:20] <Peng_> CBro2007: http://bazaar-vcs.org/Download
[22:21] <lifeless> jelmer: multiply_tests(loader.loadTestsFromModuleNames(modules), scenarios, standard_tests)
[22:22] <CBro2007> thanks Peng_
[22:22] <CBro2007> Peng_: had to add that new RPM repository in
[22:22] <CBro2007> cheers
[22:24] <thecookie> So, I searched a bit for a feature similar to svn externals.. Seemed to not be there yet, but in the works?
[22:24] <meoblast001> Peng_: thanks
[22:25] <Peng_> thecookie: Right, but it's been "in the works" for ages. I think there may have been some progress recently, or there will be soon, or something.
[22:26] <thecookie> I see
[22:26] <jelmer> lifeless, thanks
[22:31] <Kobaz> dude7064: what about #asterisknow?
[22:31] <Kobaz> er
[22:31] <Kobaz> wrong chan
[22:35] <lifeless> jelmer: (looking at any of the __init__'s changed by my patch would probably have shown you that :P)
[22:43] <lifeless> rocky: yes, revids are guids, revnos are not
[22:58]  * mwhudson waves towards brisbane
[23:03] <psynaptic> hey guys
[23:03] <psynaptic> I'm trying to login to launchpad to start using bzr for a project
[23:04] <psynaptic> I've used the bzr launchpad-login rich-freestylesystems command but it's throwing an error about ssl: http://pastie.textmate.org/private/nrknbmomsg8czxkqxz9luw
[23:04] <psynaptic> any ideas?
[23:08] <Peng_> psynaptic: My guess is that your Python installation doesn't support SSL, or you have a file called socket.py somewhere in your Python path (such as the current directory)
[23:09] <psynaptic> Peng_: is there a way I can find out for sure if my pyton lacks ssl support?
[23:09] <psynaptic> maybe I can reinstall pyton or something
[23:10] <mwhudson> psynaptic: run python and type "from socket import ssl"
[23:10] <psynaptic> ok, did that
[23:10] <psynaptic> it returned nothing, still in pyton
[23:10] <psynaptic> do I need to do something to run it?
[23:11] <mwhudson> is bzr running with the same python as you just ran?
[23:11] <mwhudson> if so, maybe Peng is right
[23:11] <psynaptic> I think so
[23:12] <mwhudson> bzr 1.1 is ages old mind
[23:12] <mwhudson> psynaptic: what command were you running when you got that traceback?
[23:13] <psynaptic> bzr launchpad-login rich-freestylesystems
[23:13] <psynaptic> I installed the latest bzr about half an hour ago
[23:16] <Peng_> psynaptic: Whether or not you installed it, what you ran was 1.1, while the current version is 1.12.
[23:16] <psynaptic> oh
[23:17] <psynaptic> just doing an update of my packages
[23:31] <psynaptic> I tried to install bzr again from macports but I get some errors: http://pastie.textmate.org/private/whynuza6yxkvemmxx12vq
[23:35] <thrope> I have a bzr-svn question  - I have two diverged branches, craeted with bzr with nothing to do with svn... they have a shared history up to revision 24 or so. I would like to push these both into a svn repository... is there a way to push them while keeping the branch relationship in svn terms (ie have one as a copy of the other)
[23:35] <thrope> i am thinking of somehow pushing up to where they diverge to svn, copying it in svn, then somehow checking out the two copies of this and rebasing the branches onto that so I can push back the diverged changes
[23:36] <thrope> but I'm having some trouble getting it to work
[23:37] <thrope> would appreciate any pointers on how to do this
[23:37] <thrope> do I need to set up an svn layout so it knows where to look for branches?
[23:41] <lifeless> vila: your desire is granted; see the commits list, give it a spin
[23:43] <jfroy> vila: FYI I submitted the patch to AuthConfig early last week, quietly sitting in the PQM
[23:50] <vila> jfroy: You mean sitting in BB, but yes I saw it, it needs a couple of tweaks, I'll review that more formally today
[23:51] <jfroy> I also have a quick patch which I may submit to fill in a default port number when None is passed, based on the scheme.
[23:51] <jfroy> Using socket.getservbyname
[23:51] <psynaptic> thanks anyway guys]
[23:53] <vila> jfroy: We tried using a default port by scheme in the past but that wasn't worth the effort (and there was hard to address problems trying to do that), why do you want or even need that ?
[23:54] <jfroy> the keychain store needs an explicit port number
[23:54] <lifeless> jfroy: adding a port to urls when one wasn't provided is a bad idea
[23:54] <jfroy> to construct the service string
[23:54] <jfroy> lifeless: I disagree, protocols have well-established default ports
[23:55] <vila> jfroy: then do what needs to be done locally in your keychain wrapper
[23:55] <jfroy> Ideally, the port should be provided by the transport layer
[23:56] <jfroy> vila: there are issues beyond the keychain provider
[23:57] <vila> jfroy: may be, but they are also issues trying to make that work in bzrlib and the python modules involved and the libraries involved, pushing the problem there, where it was tried and failed, will not help
[23:57] <jfroy> for example, specifying port=80 in a auth. config section for an http scheme, then running a command with an http URL without an explicit port number (which is a perfectly valid and expected use case) will yield no section match, because None != 80.
[23:57] <mwhudson> if bzr started faster, i'd sure use uncommit more :)
[23:57] <mwhudson> bzr ci -m <some typo>[ret]C-c is quite common for me...
[23:58] <jfroy> And that is just a very poor, unexpected behavior.
[23:59] <mwhudson> jfroy: if you manage to change things so that bzr+ssh ignores "Port <num>" directives in my .ssh/config then the masses will rise up with pitchforks &c
[23:59] <vila> jfroy: ^ see ?
[23:59] <jfroy> mwhudson: not speaking about that at all :p
[23:59] <mwhudson> (this was the problem before)
[23:59] <vila> jfroy: Yes you *are* speaking about that